Technological difficulties associated with high-frequency (>20 MHz) transducer arrays have prevented wide adoption of high-resolution imaging systems for many decades. Recently, high-frequency arrays and a supporting system operating in the 20 - 70 MHz range have been commercially developed for small animal imaging. However, the extremely expensive cost of this system (>$500 k) and associated transducer probes (~$20 k) is still a huge barrier for widespread use of this preclinical tool as well as the expansion of array-based high-resolution imaging to clinical applications in ophthalmology, dermatology, and cardiovascular medicine. The major hurdles for the development of high-frequency arrays have been related to the conventional manufacturing techniques used for piezoelectric transducers: 1) The thickness of the piezoelectric crystal needs to be tens of microns for high-frequency operation. Using conventional techniques such as lapping and polishing it is very difficult to thin crystals down to required thicknesses. 2) The pitch of the array should be on the order of a wavelength (~50 m for 30 MHz). When the crystal is diced using a standard 10 to 15-?m blade, significant portion of the active area is wasted for element separation. Therefore, methods such as plasma etching or laser machining should be employed for dicing the crystal. 3) Depending on how the kerfs are filled, increased cross coupling could be observed between elements. 4) The small element size results in high electrical impedance and makes it more difficult to drive the array using external electronics. Capacitive micro machined ultrasonic transducer (CMUT) arrays have demonstrated over the last decade that they hold a great promise for the implementation of high-frequency arrays: 1) The frequency of operation of CMUTs is set by the width and thickness of a thin vibrating plate, which can be precisely defined with sub-micron features to enable efficient high-frequency transduction of ultrasound. 2) By lithographic patterning, densely placed array elements can be isolated from each other with no need for dicing. 3) It has been experimentally demonstrated that crosstalk in neighboring CMUT elements can be as low as -39 dB. 4) Electronic circuits can be conveniently integrated with transducer arrays on the same substrate or by chip-to-chip bonding to achieve a high signal quality, low noise, and wide bandwidth. CMUTs have the potential to further extend the frequency of operation, enable high-frequency 2-D arrays and arrays with other geometries, and lower the cost of array manufacture by taking advantage of batch microfabrication. To demonstrate high-frequency wideband elevation-focused CMUTs we have identified two specific aims: Specific Aim 1: Design, implement, and test 256-element, 1-D linear CMUT arrays operating at 40 MHz and 60 MHz center frequencies with a fractional bandwidth greater than 100%. Specific Aim 2: Develop a process to curve 1-D linear arrays on thin substrates in the elevation direction and demonstrate elevation focusing without using a glossy lens in front of the array.